In this article I seek to explain the microfoundations of patronage politics in the developing world. Two distinct approaches have evolved in the literature. One puts emphasis on the demand side, arguing that patronage persists because poor voters tend to desire individualistic goods over policy. The other focuses on the supply side: few politicians offer programmatic policy, so voters have no alternative but to vote for the politicians who distribute patronage. In this study I test those competing theories using original data from Jakarta, Indonesia. I find evidence supporting the demandside theory: when both patronage and policy are offered, poor, lesseducated voters tend to demand patronage, such as jobs and money, over national programs like free education and universal health care, whereas well-off, better-educated voters tend to prefer the national policies. However, the study also reveals that demands for patronage are affected by level of participation in politics: those who voted in previous elections and those who affiliate with a political party are more likely to demand patronage. This microfoundational evidence helps to explain the persistence of patronage politics in places of widespread poverty.